1.3.1 Labeled Bracketing
How the constituents of a phrase combine can be shown via a representation that is much used in linguistics, a labeled bracketing. We introduce “labeled bracketing” here because our preferred representation, the phrase marker, evolves nicely from it.
Construction of the labeled bracketing is straightforward. Each sequence of constituents making up a phrase is flanked by paired brackets, and then a label specifying the kind of phrase is placed immediately within the leading bracket. To promote readability and minimize clutter, our illustra-tions of bracketing will be in English, with the segments in italics.
Let us construct the labeled bracketing for the three-segment word וֹי ְר ִפּ ִמ, from + fruit + its.
Consider first the common noun fruit. The labeled bracketing for fruit is [N fruit], where N indi-cates that this segment is a noun. Relying on our recollections of grade-school grammar, we assert that fruit and its may next be combined to form the larger noun phrase whose structure is [NP [N fruit] [pron its]], where “pron” indicates that its is a pronoun. The syntactic structure of the full prepositional phrase—abbreviated PP—is then represented as in (1.1).
(1.1) [PP [prep from] [NP [N fruit] [pron its]]]
This bracketing captures the syntactic structure of the prepositional phrase, but correctly pairing the left brackets with their corresponding right brackets can be tedious, especially when the constitu-ents become lengthy. We will overcome this limitation in the next section.
The decoding of labeled brackets only becomes more arduous as more complex constituents are analyzed. For example, consider this six-word clause from 2 Kgs 18:8:
ה ָזּ ַע־ד ַע םי ִתּ ְשׁ ִל ְפּ־ת ֶא ה ָכּ ִה־אוּה We translate the six words literally as:
He + he-hit + [nota acc.] + Philistines + until + Gazah.
The labels needed for representing the structure of this clause are:
Label Constituent Name
In terms of these labels, the labeled bracketing for this simple clause is as in (1.2).
(1.2) [S [pron he][VP [V he-hit][PP [prep [nota acc.]][PN Philistines]][PP [prep until][PN Gazah]]]]
This is not easy to decipher. Its structure can be made clearer by introducing judicious indenting as is shown in (1.3), that is, by making an indented list.
(1.3) [S
In this representation, we keep the words and their part-of-speech labels together with their enclos-ing brackets on the same vertical line. The constituents in the column immediately to the right of a given constituent down to its closing bracket are called its immediate constituents (ICs). Thus, for example, the immediate constituents of S (the sentence) are a pron (pronoun) and a VP (verb phrase). The VP has three ICs: V (the sentence’s verb) plus two PPs (prepositional phrases). The first PP has two ICs: the nota accusativi (object marker) and a proper noun. Even this representa-tion is awkward to read. Enter the phrase marker. . . .
1.3.2 The Phrase Marker
Suppose that we link each constituent with each of its immediate constituents via an arrow from the former to the latter and that we do away with the now superfluous brackets. Then the representa-tion in (1.4), mathematically equivalent to a labeled bracketing, is obtained. This is a phrase marker.
31. For now, so as to parallel standard works on syntax, we have chosen to name the most inclusive constituent the sentence. However, as our exposition develops below, our major unit of syntactic structure will be termed the clause.
(1.4)
Unlike a labeled bracketing, a phrase marker is easily understood. For example, in phrase marker (1.4), we readily see that the proper noun ‘Gazah’ combines with a preposition (‘until’) to form a prepositional phrase (PP = ‘until Gazah’), which is one of three immediate constituents of the sentence’s verb phrase (VP = ‘he-hit [nota acc.] Philistines until Gazah’).
Mathematically speaking, a phrase marker is a tree. If the phrase marker is counter-rotated by ninety degrees so that the S is at the bottom (the tree root) and the words are at the top (the tree leaves), then the reason for calling the structure a tree will be obvious. In linguistics textbooks, phrase markers are invariably displayed with the root at the top and the leaves at the bottom (i.e., the tree is upside down). We opt for placing the tree on its side with its root to the left so that long clauses can be displayed down the page or screen rather than across.
1.3.3 The Enhanced Phrase Marker 1.3.3.1 Enhancements Made
Being fashioned in the traditional way, phrase markers of the sort presented above show how larger and larger constituents are built up until they finally combine to show the structure of entire clauses.
As we developed our syntactic representation, we asked what additional information might be incorporated in a phrase marker to enhance its usefulness for research. One interesting and well-developed syntactic theory, lexical functional grammar, argues that a complete presentation of syn-tax must supply both a picture of the hierarchical makeup of clauses and a quite separate catalog of essential information regarding grammatical functions. “[Functional]-structure is composed of attributes (features and functions) and their values.” 32 Included in functional-structure is informa-tion specifying the subject, direct object, tense, and so on, associated with a clause.
Much functional-structure information can be straightforwardly grafted onto traditional phrase markers, as can information regarding the semantic characteristics of the constituents. To accom-plish this enrichment (and make our phrase markers more readable), we make these changes and enhancements to traditional phrase markers:
1. Add the Hebrew text to the left of the English glosses.
2. Replace S (sentence) by cl (clause) in the root node at the far left of the phrase markers.
3. Delete the VP (verb phrase) constituent, joining its constituents directly to the root node.
4. Then, to the immediate right of the root node, insert a set of function-specifying labels, short yet informative (e.g., dir obj) rather than acronymic (e.g., do), that provide the names
32. Y. N. Falk, Lexical-Functional Grammar: An Introduction to Parallel Constraint-Based Syntax (Stanford, CA:
CSLI, 2001) 16.
of the clause immediate constituents (CICs).
5. Enrich the label information by introducing a two-line format. The upper line describes the form or function of the constituent. When form is being described, explicit semantic information is included. The lower line gives the licensing relation justifying the
assembling of the constituent.
The resulting structures might properly be termed enhanced phrase markers, but for brevity we will simply continue to call them phrase markers.
Transforming the traditional phrase marker (1.4) yields the (enhanced) phrase marker (1.5). 34 (1.5)
The innovations given in the previous list are realized in this phrase marker as follows:
Item #1: Hebrew text. Self evident.
Item #2: “Clause” as Root Constituent. In a simple substitution, cl replaces S. Along with the newly minted cl in the upper line of the root node label, a new shortened form has appeared in the lower line: oblq. This is short for obliqueness, a topic taken up in §1.3.3.3.
Item #3: Deletion of the Verb Phrase. In most modern textbooks on syntax, one of the fixed points is that the root label (be it symbolized by S or cl) in the phrase marker for a declarative sen-tence always has two and only two immediate constituents, NP and VP. The former functions as the clause’s subject, while the latter functions as the clause’s predicate. This is the orthodox view.
Practitioners of lexical functional grammar and others, however, suggest that holding this binary view of sentence structure is counterproductive when certain non-English-like languages, so-called nonconfigurational or fluid languages, are being investigated. We address this topic in §7.2 and
§7.3.3. We trust that until then readers will, however uneasily, accept our jettisoning of the use of VPs for Biblical Hebrew.
Items #4 and #5: CIC Insertion and Label Format. The inclusion of function-specifying la-bels is a major innovation. This clause has the four clause immediate constituents (CICs) named in the upper lines of the CIC labels. Their expanded names are as follows:
33. Clause immediate constituents (CICs) are the major constituents in clauses. They are treated at great length in chaps. 9–16. Clause immediate constituents have also been referred to as clause-rank constituents in John H. Connolly, Constituent Order in Functional Grammar: Synchronic and Diachronic Perspectives (Berlin: Foris, 1991) 28. The eas-ily pronounced acronym CIC (“kick”) is our innovation.
34. This and all of the subsequent phrase markers in this book have been exported from the Andersen-Forbes Phrase Marker Analysis of the Hebrew Bible, as displayed by the Libronix Digital Library System, Logos Research Systems.
CIC Label Expanded Name
sbj subject
vb finite verb, one kind of predicator dir obj direct object
mvt aim movement aim
The subject is often said to be “what in the clause is already known” (its theme). Traditionally, it is the doer of the action in a clause. The predicator is not the traditional predicate. 35 Rather, it is a verbal or quasiverbal constituent that specifies an equivalence, a state, an activity, or a process.
In the present example, the predicator is a finite verb. The direct object is that which is affected by the action specified by the predicator. It is the “undergoer” of the action of the predicator. In Bibli-cal Hebrew, the direct object is often marked by the nota accusativi, as it is here. The movement aim is the destination of the activity specified in the clause. The lower lines of the CIC labels each read gram. This indicates that a CIC classification has been assigned on the basis of grammatical knowledge. The significance of this label is discussed in §1.3.3.3.
1.3.3.2 Construction-Identifier Label Formats
The construction-specifying upper lines in the labels on non-CIC nodes come in two flavors:
one for prepositional phrases and one for noun phrases.
Prepositional Phrase Construction Format. Our example clause has two other labels to the right of the four CIC labels just explained. These non-CIC labels are in the prepositional phrase format. The upper line in the label on the first prepositional phrase constituent ‘[nota acc.] + Phi-listines’ reads obj mk+humn, meaning that the prepositional phrase consists of the object-marking preposition plus (+) a noun having human semantics. The upper line in the label on the preposi-tional phrase “until + Gazah” reads until+geog, meaning that the prepositional phrase consists of the preposition “until” plus a noun having location / geographic semantics. In both labels, the lower line of the label reads prep, indicating that we are dealing with a prepositional phrase gathered to-gether (“licensed”) because of the presence of a preposition.
Noun Phrase Construction Format. The noun phrase format is different. Consider phrase-marker fragment (1.6) from 2 Chr 29:32.
(1.6)
35. The term predicator has a long history. For example, in 1961, Halliday used it as we do: M. A. K. Halliday, Categories in the Theory of Grammar (Indianapolis, IN: Bobs-Merrill, 1961) 257.
The constituent is a noun phrase with animal semantics (n:anml) produced by juxtaposing (juxt) its three subconstituents. Each of the three subconstituents is itself a noun phrase with animal semantics (n:anml) produced by modifying (mod) the first segment in each phrase with the next segment.
The upper line reads n:anml, indicating that the constituent formed is a noun (phrase) further specified (:) as involving animal semantics.
1.3.3.3 The Licensing Relations of Grammar and Obliqueness
When we say that the assembling of a prepositional phrase is licensed by the presence of an initial preposition or that the assembling of a construct phrase is licensed by the presence of an ini-tial substantive in the construct state, the underlying grammatical mechanisms being described are clear. But when we say that the assigning of some label to a clause immediate constituent (CIC) is licensed by grammar [gram] or that the assembling of a clause is licensed by obliqueness [oblq], the underlying grammatical mechanisms are unspecified.
Given our adoption of a two-line format for the node labels in our phrase markers, we had to assign some licensing relations to justify the classification of CICs and the assembling of clauses;
but which ones and, more importantly, with what meanings? We might have simply installed the label dummy, explicitly signaling our initial ignorance of the what, how, and why of CIC clas-sification and clause assemblage. Instead, we opted for grammar and obliqueness, allowing us to search for each licensing relation independently. The relations grammar and obliqueness stand for the very disparate and still obscure mechanisms that provide the grounds for classifying CICs and assembling them into clauses. Here is the critical point for our work:
Gathering the data needed to begin working out the grammatical mechanisms underlying
“grammar” and “obliqueness” is the focus of the latter two-thirds of this book.